J. Neurochem. (2012) 120, 765–778.
Exposure to psychostimulants results in neuroadaptive changes of the mesencephalic dopaminergic system including morphological reorganization of dopaminergic neurons. Increased dendrite arborization and soma area were previously observed in primary cultures of mesencephalic dopaminergic neurons after 3-day exposure to dopamine agonists via activation of D3 autoreceptors (D3R). In this work, we showed that cocaine significantly increased dendritic arborization and soma area of dopaminergic neurons from E12.5 mouse embryos by activating phosphorylation of extracellular signal-regulated kinase (ERK) and thymoma viral proto-oncogene (Akt). These effects were dependent on functional D3R expression because cocaine did not produce morphological changes or ERK/Akt phosphorylation neither in primary cultures of D3R mutant mice nor following pharmacologic blockade with D3R antagonists SB-277011-A and S-33084. Cocaine effects on morphology and ERK/Akt phosphorylation were inhibited by pre-incubation with the phosphatidylinositol 3-kinase inhibitor LY294002. These observations were corroborated in vivo by morphometrical assessment of mesencephalic dopaminergic neurons of P1 newborns exposed to cocaine from E12.5 to E16.5. Cocaine increased the soma area of wild-type but not of D3R mutant mice, supporting the translational value of primary culture. These findings indicate a direct involvement of D3R and ERK/Akt pathways as critical mediators of cocaine-induced structural plasticity, suggesting their involvement in psychostimulant addiction.
thymoma viral proto-oncogene
brain-derived neurotrophic factor
bovine serum albumin
D3R null mutant mice
extracellular signal-regulated kinase
phosphorylated thymoma viral proto-oncogene
phosphorylated extracellular signal-regulated kinase
protein kinase C
ventral tegmental area
The evidence that cocaine addiction is associated with neuroadaptive changes of mesolimbic dopaminergic pathway is well established (Nestler et al. 1993; Koob et al. 1998; Thomas et al. 2008). While several studies addressed gene expression and neurochemical changes, fewer works describe changes of cell morphology, phenomena also defined as structural plasticity. These events are of potential interest because representative of long-lasting reorganization of synaptic input and signal processing underling the chronic features of addiction (Nestler et al. 1993; Robinson and Kolb 2004; Russo et al. 2010; Stuber et al. 2010). In rodents, exposure to cocaine or amphetamine produces structural changes of dendrite morphology of the dopaminoceptive medium spiny neurons of the nucleus accumbens and in pyramidal neurons of the prefrontal cortex (Robinson and Kolb 1999, 2004; Lee et al. 2005). In ventral mesencephalon, significant outgrowth of dendrites of dopaminergic neurons was observed after repeated exposure to amphetamine (Mueller et al. 2006), while a single dose of cocaine increased the number of synaptic spines in dopaminergic neurons of the mesencephalic ventral tegmental area (VTA) (Sarti et al. 2007). Prenatal exposure to cocaine can also result in long-term changes in the mesencephalic dopaminergic system of the adult offspring (Fang and Ronnekleiv 1999; Glatt et al. 2000).
The molecular mechanisms of structural plasticity to cocaine and other psychostimulants are only partially understood. A parsimonious hypothesis asserts that the extracellular increase of dopamine (DA) produced by these drugs is the critical factor leading to morphological changes. These changes may lead to excessive activation of DA D1, D2 and D3 receptors (D1R, D2R and D3R) in the nucleus accumbens or of D2R and/or D3R auto-receptors expressed in the mesencephalic dopaminergic neurons (Missale et al. 1998). A relevant role for D3R was suggested because its expression and function is enhanced by cocaine exposure (Caine and Koob 1993; Le Foll et al. 2002; Neisewander et al. 2004; Heidbreder et al. 2005). Recent findings implicate D3R in structural neuroplasticity and regeneration of mesencephalic dopaminergic neurons both in vivo (Joyce and Millan 2007) and in vitro (Collo et al. 2008). In the latter study, we observed increases of dendrite arborization and soma size of mesencephalic dopaminergic neurons in mouse primary cultures following treatment with the preferential D3R agonists and amphetamine. These effects were dependent on phosphorylation of the extracellular signal-regulated kinase (ERK) pathway (Collo et al. 2008). Interestingly, ERK and the thymoma viral proto-oncogene (Akt) pathways play a critical role for dendrite outgrowth in hippocampal neurons (Jaworski et al. 2005; Kumar et al. 2005) and for soma size in dopaminergic neurons of the substantia nigra (SN) (Domanskyi et al. 2011; Kim et al. 2011).
In the present work, we investigated the effects of cocaine on structural plasticity of mesencephalic dopaminergic neurons. The roles of D3R and of ERK/Akt pathways were explored in primary cultures from wild-type or D3R knockout mice (D3KO) or using pharmacologic antagonists. The in vivo relevance of these observations was obtained by assessing structural plasticity in SN and VTA of newborn mice exposed to cocaine in utero at the same embryonic stage used for the in vitro cultures.
Materials and methods
Cocaine was purchased from Sigma (St Louis, MO, USA); 7-hydroxy-N,N-di-propyl-2-aminotetralin (7-OH-DPAT), sulpiride, SCH23990 and LY294002 from Tocris Bioscience (Bristol, UK); brain derived neurotrophic factor (BDNF) from Alomone Labs Ltd. (Jerusalem, Israel); SB-277011-A was a gift from C. Heidbreder, GlaxoSmithKline (Verona, Italy); S-33084 was provided by Institut Servier (Paris, France). Cocaine, 7-OH-DPAT, SCH23990, S-33084, and BDNF were dissolved in water; SB-277011-A and LY294002 were dissolved in dimethylsulfoxide (> 1 : 3000 final dilution), while sulpiride was dissolved in ethanol (> 1 : 2000 final dilution), all as recommended by producer. For each control treatment the vehicle was prepared with the respective dilution of dimethylsulfoxide or ethanol.
CD1 mice were obtained from Charles-River Laboratories (Italy). Mice genetically deprived of dopamine D3R (D3KO) and their singenic wild-type were obtained from the Jackson Laboratory (B6.129S4-Drd3Tm1Dac/J) (Accili et al. 1996). Homozygous mice were generated on a pure genetic background and compared with singenic wild-type mice. The genotype of D3KO and singenic mice was assessed as described in Accili et al. (1996). Breeding of all mice was performed to achieve timed pregnancy with the accuracy of ± 0.5 days. The embryonic day (E) was determined by considering the day of insemination (determined by vaginal plug) as day E 0.5. All mice were cared and killed in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC) and in line with Italian law.
Primary mesencephalic cultures
Primary mesencephalic cultures were prepared as described in Collo et al. (2008). Briefly, ventral mesencephalon tissues were dissected from E12.5 mouse embryos and collected in ice-cold Hank’s buffered salt solution (EuroClone, Milan, Italy). Tissues were transferred in Accumax (Sigma Aldrich), mechanically dissociated at 22°C and resuspended in Neurobasal medium (Gibco-Invitrogen, Carlsbad, CA, USA) containing 2 mM glutamine (EuroClone) and B27 supplement (Gibco-Invitrogen). Cells were counted and seeded on poly-d-lysine/laminin (Sigma-Aldrich) -coated coverslides (5 × 104/mL), or 12-well plates (5 × 105/mL), and cultured at 37°C in a humidified atmosphere of 5% CO2 and 95% air. No fetal bovine serum was added to the cultures. Fifty percent of the medium was changed every 2 days until treatment. Under these conditions, the cultures contained > 99.5% neurons and < 0.5% astrocytes, as assessed by microtubule-associated protein 2 and glial fibrillary acidic protein immunostaining, respectively (data not shown). Pharmacological treatments were conducted in vitro after 5 days from seeding, time required for primary culture stabilization.
In vivo prenatal cocaine treatment
Pregnant D3KO and singenic wild-type mice were individually housed in a climate controlled room on a 12/12-h light/dark cycle with ad libitum access to food and water. Treatments were performed during the light period (9:00 am to 6:00 pm). All mice were weighed daily starting at day 10. At day 12.5, pregnant mice weighting 25–30 g were administered intraperitoneally (i.p.) with two daily dose of 20 mg/kg cocaine or saline as described by Crozatier et al. (2003). In this article, this regimen resulted in enhanced stereotypes to cocaine challenges in the offspring as young adults, indicating neuroadaptation. Treatments were repeated for 5 days, between E12.5 to E17. After each injection, animals were placed for 30 min in a clean cage and inspected for locomotor activity. After delivery, P1 newborn mice were killed and the brains were removed and processed for immunohistochemistry.
Immunocytochemistry and immunofluorescence in primary cultures
Immunocytochemistry and immunofluorescence were performed as described in Collo et al. (2008). Briefly, mesencephalic neurons were fixed in phosphate-buffered saline (PBS) containing 3% paraformaldehyde (Sigma-Aldrich)/3% sucrose (Sigma Aldrich), blocked in PBS-containing 0.1% Triton (Promega, Madison, WI, USA), 5% normal goat serum (Jackson ImmunoResearch, West Grove, PA, USA), 1% bovine serum albumin (BSA; Sigma Aldrich) and incubated overnight at 4°C with a rabbit polyclonal anti-tyrosine hydroxylase (TH) antibody (1 : 500; Santa Cruz Biotechnology, Santa Cruz, CA, USA), followed by a biotinylated goat anti-rabbit antibody (Jackson ImmunoResearch) (30 min at 22°C) and a final incubation with the ABComplex/horseradish peroxidase detection system (Vector Laboratories, Burlingame, CA, USA). Color development was achieved by addition of 3,3′diaminobenzidine (Sigma Aldrich) for 45 s. In immunofluorescence, the dopaminergic neurons were identified using either anti-TH or anti-DAT antibodies, whose co-distribution was previously tested. To study the expression of phosphorylated ERK (p-ERK) or Akt (p-Akt) in dopaminergic neurons, mesencephalic cultures were incubated with either a rabbit polyclonal antibody anti-TH (1 : 800; Santa Cruz Biotechnology) or a rat monoclonal antibody anti-DA transporter (1 : 400; DAT) (Santa Cruz Biotechnology), together with a mouse monoclonal antibody to p-ERK (1 : 400; Cell Signaling Technology, Beverly, MA, USA) or a rabbit polyclonal antibody to p-Akt (1 : 200; Cell Signaling Technology), respectively. DyLightTM 488- and Cy3-conjugated secondary antibodies (Jackson Immuno Research) were used. Coverslides were mounted with Vectashield Mounting Medium (Vector Laboratories). As negative control, cultured neurons were incubated with the secondary antibodies only. Immunofluorescence was visualized with an Olympus IX51 microscope (Hamburg, Germany). Each experiment was repeated three times and all dopaminergic neurons present in two coverslides for each treatment condition were assessed.
Immunohistochemistry for brain sections
Newborn mice were killed at P1 and brain removed, fixed overnight in PBS 4% paraformaldehyde, cryoprotected in 20% sucrose and rapidly frozen by immersion in isopentane on dry ice. For each brain, a complete set of coronal sections was cut through the SN and VTA at 30 μm. Sections were mounted on slides, blocked and permeabilized with PBS, 5% BSA (Sigma-Aldrich), 0.1% Triton (Sigma-Aldrich) for 30 min at 22°C. They were incubated with the rabbit anti-TH antibody (1 : 500; Santa Cruz) diluted in PBS, 1% BSA, overnight at 4°C. Then they were washed and incubated with a biotinylated goat anti-rabbit antibody (Jackson ImmunoResearch) diluted in PBS, 1% BSA for 30 min at 22°C, followed by avidin-biotinylated horseradish peroxidase complex (ABC; Vector Laboratories) and development with DAB.
Computer-assisted morphological analysis
Digital images from primary neuronal cultures or mesencephalic brain tissues were acquired with the Olympus IX51 microscope connected to an Olympus digital camera and a PC. Morphometric measurements were performed by a blinded examiner on digitalized images using Image-Pro Plus software (Media Cybernetics, Bethesda, MD, USA). For in vitro primary neuronal culture, the morphological indicators of structural plasticity were: (i) the maximal dendrite length, (ii) the number of primary dendrites and (iii) the soma area (Schmidt et al. 1996). Maximal dendrite length was defined as the distance from the soma (hillock base) to the tip of the longest dendrite for each neuron; dendrites shorter than 20 μm were excluded from the analysis. Primary dendrites were defined as those directly stemming from the soma. Soma area was assessed by measuring the surface (μm2) included by the external perimeter drawn on the cell membrane of neurons identified by TH-IR staining. Two coverslides per treatment groups were examined so to obtain measurements from at least 30 neurons. For ex vivo studies computer assisted morphometry was performed using procedures modified from Russo et al. (2007). The effects of cocaine on newborn mice were assessed by measuring in each brain the soma area of 90–120 dopaminergic neurons at various rostrocaudal levels of SN and VTA, respectively. The same rostrocaudal levels were considered in each treatment group so to keep anatomical variance partially controlled. In a typical experiment 8 newborn mice from three different mothers per group were included. The soma area of dopaminergic neurons on histological preparations was measured as described for TH-IR primary neurons in cultures.
Western blotting was performed at different time points (1–10 min) following challenge with cocaine (10 μM) and/or after pre-treatments with either D3R antagonist SB-277011-A (50 nM) or phosphatidylinositol-3-kinase (PI3K) antagonist LY294002 (10 μM). For studies in D3KO and singenic wild-type mice, challenges were produced with cocaine (10 μM) and 7-OH-DPAT (50 nM). Mesencephalic cultures were washed with ice-cold PBS and lysed in a buffer containing 50 mM Tris (pH 7.4), 150 mM NaCl, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulphate, 1% Igepal, 1 mM polymethanesulphonyl fluoride, and complete protease inhibitors (Roche Diagnostics, Mannhein, Germany) for 10 min in ice. Neurons were homogenized with a probe- type sonicator on ice, and centrifugated at 13 000 g for 10 min at 4°C. The protein concentration was measured with a DC-protein assay (Bio-Rad, Hercules, CA, USA). Twenty micrograms of total proteins were resolved by 10% sodium dodecyl sulphate polyacrylamide gel electrophoresis and blotted onto a PVDF membrane (Immobilon-P; Millipore, Billerica, MA, USA). Membranes were blocked in TBS/T (25 mM Tris base, 137 mM NaCl, 2.7 mM KCl and 0.1% Tween-20, pH 7.6), 5% dry milk, and incubated overnight with the primary antibody diluted in 5% milk in TBS/T. The primary antibodies used were: anti-p-ERK (1 : 1000) and anti-ERK (1 : 1000) mouse monoclonal antibodies (both from Santa Cruz Biotechnology), anti-pAkt (1 : 2000) and anti-Akt (1 : 1000) rabbit polyclonal antibodies (both from Cell Signaling Technology) and anti-TH (1 : 1000) rabbit polyclonal antibody (Millipore). In a typical experiment, the same membrane was processed in the following order: incubation with anti-pERK antibody, stripping, incubation with anti-ERK antibody, incubation with anti-p-Akt antibody, stripping, incubation with anti-Akt antibody, and incubation with anti-TH antibody. At each step, blots were washed and incubated for 90 min with appropriate horseradish peroxidase-conjugated secondary antibodies (goat anti-rabbit or goat anti-mouse; Santa Cruz Biotechnology), and developed using ECL (Amersham Biosciences, Milan, Italy). Specific bands were analyzed by densitometric scanning of exposed film using Gel-pro analyzer software (Media Cybernetics). In each experiment, the specific signal of p-ERK or p-Akt proteins was normalised to the corresponding ERK or Akt signals, respectively, and then to the level of TH measured in the same preparation.
Assay of [3H]DA uptake
[3H]DA uptake was measured as described in Collo et al. (2008). To eliminate any residual cocaine and to prevent any effect caused by its direct binding to the DA transporter (DAT), mesencephalic cultures were washed and incubated with fresh medium 16 h before performing the [3H]DA uptake test. At the moment of the test the mesencephalic neurons were washed twice with Krebs–Ringer buffer containing 16 mM sodium phosphate, 119 mM NaCl, 4.7 mM KCl, 1.8 mM CaCl2, 1.2 mM MgSO4, 1.3 mM EDTA, and 5.6 mM glucose (pH 7.4), and incubated in Krebs–Ringer buffer containing 1 μM [3H]DA (45.0 Ci/mmol; Amersham Biosciences) for 15 min at 37°C. At the end of the incubation, the cultures were washed three times with ice-cold Krebs–Ringer buffer, and the cells were collected in 1 M NaOH. The radioactivity was counted in a Beckman liquid scintillation counter.
Data were expressed as mean ± standard error of the mean (SEM) if not stated otherwise. Significant differences from control conditions were determined using either one- or two-way analysis of variance (anova) followed by a posteriori Bonferroni’s test for multiple comparisons provided by GraphPad Prism version 4.00 for Windows (GraphPad Software, San Diego, CA, USA).
Cocaine-induced structural plasticity in primary cultures of mesencephalic dopaminergic neurons of the mouse
TH-IR dopaminergic neurons were visualized, as represented in Fig. 1a and b. The time-dependency of dendrite outgrowth of dopaminergic neurons, measured as maximal dendrite length and number of primary dendrites, and of soma size, measured as area, was assessed by studying the effects of various doses of cocaine (0.1, 1 and 10 μM) or vehicle at two different time points, 24 and 72 h. Data were analyzed using a two-way anova. The maximal length of dendrites showed a significant treatment effects [F(3,232) = 46.36; p < 0.0001] with significant difference from vehicle already at 1 μM and stronger effect at 10 μM (Bonferroni’s test, p < 0.001 vs. vehicle), while time effect and interaction were not significant (Fig. 1c). The analysis of the number of primary dendrites showed a similar cocaine dose-related increase [F(3,232) = 60.63, p < 0.0001], with significant differences at 1 and 10 μM dose (Bonferroni’s test, p < 0.001 vs. vehicle) and a stronger significant effect at 10 μM (Bonferroni’s test, p < 0.001 10 μM vs. 1 μM) (Fig. 1d). The main effect of time was also significant [F(1,232) = 6.70; p < 0.02] with the strongest signal at 72 h (Bonferroni’s test, p < 0.01), while ‘treatment × time’ interaction was not significant. Differently, the soma area showed significant ‘treatment × time’ interaction [F(3,224) = 6.26; p < 0.0005], as well as treatment [F(3,232) = 25.33; p < 0.0001] and time effect [F(1,232) = 17,02; p < 0.0001]. Post hoc analysis indicated that the significant interaction was caused by a modest increase during the first 24 h and a concomitant strong growth at 72 h, mostly detected at the dose of 10 μM (Bonferroni’s test, p < 0.001 vs. vehicle) (Fig. 1e). We concluded that the best test parameters for pharmacological studies were: a cocaine dose of 10 μM and an incubation time of 72 h.
Cocaine-induced structural plasticity in primary mesencephalic cultures: pharmacological studies with DA antagonists
The involvement of DA receptors in dendrite outgrowth of mesencephalic dopaminergic neurons exposed to cocaine (10 μM) for 72 h was initially tested using the D2/D3R antagonist sulpiride (5 μM) and the D1R antagonist SCH23390 (1 μM) added to the culture 20 min before cocaine. On maximal dendrite length two-way anova‘cocaine × antagonist’ indicated significant cocaine effect [F(1,174) = 155,10; p < 0.0001], significant antagonist effect [F(2,174) = 35,63; p < 0.0001], and significant ‘cocaine × antagonist’ interaction [F(2,168) = 35,83; p < 0.0001] (Fig. 2a). Post hoc analysis indicated that sulpiride, but not SCH23390, significantly reduced maximal dendrite length to vehicle level (Bonferroni’s test, p < 0.001). On number of primary dendrites, two-way anova indicated significant cocaine effect [F(1,174) = 111,10; p < 0.0001], significant antagonist effect [F(2,174) = 27,66; p < 0.0001], and significant ‘cocaine × antagonist’ interaction [F(2,168) = 26,64; p < 0.0001] (Fig. 2b). Post hoc analysis indicated that pre-treatment with sulpiride, but not SCH23390, significantly attenuated the effects of cocaine (p < 0.001). SCH23390 lack of effects was expected, because no D1R is expressed in the TH-IR neurons of rodents. Incubations with sulpiride or SCH23390 alone were ineffective.
To investigate the specific role of D3R in structural plasticity similar experiments were performed with various doses of the D3R antagonists SB-277011-A and S-33084. SB-277011-A produced a dose-dependent attenuation of cocaine induced maximal dendrite length [F(5,201) = 18,55; p < 0.0001] (Fig. 2c) and number of primary dendrites [F(5,234) = 4,35; p < 0.001] (Fig. 2d). For both axonal growth parameters, post hoc analysis showed significant effects of cocaine versus vehicle (Bonferroni’s test, p < 0.001) and a blocking effect of SB-277011-A starting at doses of 50 nM (Bonferroni’s test, p < 0.05 or lower, see Fig. 2c and d). The involvement of D3R was confirmed using a more potent and selective D3R antagonist, S-33084. Pre-treatments with S-33084 significantly antagonized the effects of cocaine on the maximal dendrite length [F(5,498) = 16,94; p < 0.0001] (Fig. 2e) and the number of primary dendrites [F(5,381) = 12,94; p < 0.0001] (Fig. 2f) at doses as low as 2 nM (p < 0.001). Exposure with 50 nM of SB-277011-A alone and 10 nM S-33084 alone did not show any effect.
To test if the presence of D3R antagonists may interfere with dendrite outgrowth response independently from a D3R selective blockade, mesencephalic cultures were exposed to SB-277011-A (50 nM) or vehicle followed by BDNF (10 ng/mL) or vehicle and incubated for 72 h. Two-way anova showed that BDNF increased maximal dendrite length [F(1,136) = 105.04; p < 0.0001] that was not blocked by pre-treatment with SB-277011-A (Fig. 2g).
The [3H]DA uptake paradigm (Fig. 2h) was used to assess the functional capacity of DAT in internalizing extracellular DA, a specific feature of dopaminergic neurons. Treatments with 10 μM cocaine produced a marked [3H]DA uptake when tested after 72 h culture as shown by significant interaction [F(1,8) = 79,68; p < 0.0001]. Pre-treatment with SB-277011-A (50 nM) significantly blocked the effects of cocaine (Bonferroni’s test, p < 0.001) to the level of vehicle treatment (Fig. 2h). Incubations with SB-277011-A alone did not affect the [3H]DA uptake.
Cocaine-induced structural plasticity in primary mesencephalic cultures from D3KO mice
Mesencephalic neurons obtained from D3KO and singenic wild-type mice were processed in separated but parallel culture wells within the same experiment. Either cocaine (10 μM), the D3R agonist 7-OH-DPAT (5 nM), or BDNF (10 ng/mL) were administered 72 h before morphological assessment to induce structural plasticity. Two-way anova was used to assess the ‘genotype × treatment’ effect. Maximal dendrite length was affected by genotype [F(1,229) = 3.84; p < 0.05] and by treatment [F(3,229) = 19.42; p < 0.0001], the interaction being not significant. Post hoc assessment showed that cultures prepared from wild-type mice responded significantly to cocaine, 7-OH-DPAT and BDNF, while cultures prepared from D3KO mice significantly responded only to BDNF (Bonferroni’s test, p < 0.001 vs. control), but not to cocaine or 7-OH-DPAT (Fig. 3a). A similar profile was observed for the number of primary dendrites, showing a significant effect of genotype [F(1,392 = 29.95; p < 0.0001], treatment [F(3,392) = 14.89; p < 0.0001], and also a significant interaction [F(3,384) = 4.37; p < 0.005] (Fig. 3b). Post hoc assessment showed that in cultures prepared from wild-type mice cocaine, BDNF and 7-OH-DPAT produced morphological changes, while cultures prepared from D3KO mice significantly responded only to BDNF (Bonferroni’s test, p < 0.001 vs. control), but not to cocaine or 7-OH-DPAT (Fig. 3b). Also soma area showed significant genotype [F(1,231 = 9.94; p < 0.002] and treatment effect [F(3,231) = 9.09; p < 0.0001], while interaction was not significant. Post hoc assessment showed that cultures prepared from wild-type mice responded significantly to cocaine, 7-OH-DPAT and BDNF, while cultures prepared from D3KO mice significantly responded only to BDNF (Bonferroni’s test, p < 0.001 vs. control), but not to cocaine or 7-OH-DPAT (Fig. 3c). This dissociation indicates that neurons carrying a D3R mutation are still capable of morphological response to neurotrophic adaptive stimuli.
Cocaine-induced structural plasticity: ex vivo studies in wild-type and D3KO mice
The translational relevance of the observed structural plasticity of dopaminergic primary neurons in vitro was investigated by exposing wild-type and D3KO mouse embryos in uterus to a regimen of repeated daily treatment with cocaine and by analyzing their soma size in VTA and SN (Fig. 4a and b). Cocaine increased the soma area of TH-IR neurons in VTA of wild-type mice but did not in the D3KO mice, as supported by highly significant genotype effects [F(1,32 = 79.53; p < 0.0001], treatment effects [F(1,32 = 39.05; p < 0.0001], and ‘genotype × treatment’ interaction [F(1,28 = 49.18; p < 0.0001] (Fig. 4c). Similar results were obtained assessing the soma area of dopaminergic neurons in SN showing a highly significant genotype effects [F(1,32) = 86.69; p < 0.0001], treatment effects [F(1,32) = 73.76; p < 0.0001] and ‘genotype × treatment’ interaction [F(1,28 = 73.76; p < 0.0001] (Fig. 4d).
Intracellular ERK and Akt pathway activation by cocaine in primary dopaminergic neurons of mesencephalic cultures
The effect of cocaine on ERK phosphorylation was investigated by challenging mesencephalic neurons with 10 μM cocaine. Western blot analysis (one-way anova) showed a rapid increase of p-ERK [F(4,25) = 7.05; p < 0.001] with a significant peak activity versus vehicle at 2 min (Bonferroni’s test, p < 0.001) (Fig. 5a and b). Incubation with SB-277011-A (50 nM) 20 min before challenge significantly attenuated the p-ERK increase [F(1,16) = 12.28; p < 0.001] (Fig. 5c and d) providing evidence of D3R involvement. No effect on p-ERK was produced by SB-277011-A when tested alone. In parallel experiments, cellular localization of p-ERK in dopaminergic neurons was investigated 2 min after cocaine exposure using double immunofluorescence. Cocaine activated p-ERK-IR in TH-IR neurons (Fig. 5e–g), while its effect was blocked by pre-treatment with SB-277011-A (Fig. 5h–k). In control experiments, incubation with SB-277011-A alone did not affect p-ERK expression (Fig. 5m–o).
The effect of cocaine on Akt phosphorylation was investigated by challenging mesencephalic neurons with 10 μM cocaine. Western blot analysis showed a rapid increase of Akt phosphorylation [F(4,14) = 16.40; p < 0.001] reaching the peak activity at 2 min exposure (Bonferroni’s test. p < 0.01) (Fig. 6a and b). Incubation with SB-277011-A (50 nM) 20 min before challenge significantly attenuated the increase of Akt phosphorylation produced by cocaine [F(1,12) = 22.78; p < 0.001] (Fig. 6c and d).
In parallel experiments, cellular localization of p-Akt in dopaminergic neurons was investigated 2 min after cocaine exposure by double immunofluorescence. As TH and p-Akt antibodies were both generated in rabbit, a rat monoclonal antibody to DAT was used instead. These experiments showed that cocaine activated the expression of p-Akt-IR in DAT-IR neurons (Fig. 6e–g), while no p-Akt-IR was found in DAT-IR neurons after pre-treatment with SB-277011-A (Fig. 6h–k). In control experiments incubation with SB-277011-A alone did not affect p-Akt expression (Fig. 6m–o).
Lack of activation of ERK and Akt pathways induced by cocaine and D3R agonist 7-OH-DPAT in primary mesencephalic cultures from D3KO mice
Mesencephalic neurons obtained from embryos of D3KO and singenic wild-type mice were processed in separated cultures within the same experiment. The effect of cocaine (10 μM) or 7-OH-DPAT (50 nM) on ERK phosphorylation was investigated using western blot analysis 2 min after treatment. Cocaine and 7-OH-DPAT significantly increased p-ERK in wild-type mice but not in D3KO mice as indicated by the two-way anova highly significant genotype effect [F(1,18) = 165.8, p < 0.0001], treatment effect [F(2,18) = 35.38, p < 0.0001], and ‘treatment × genotype’ interaction [F(1,14) = 64.38; p < 0.0001] (Fig. 7a). Similar response profiles were observed by studying p-Akt in mesencephalic neurons. Increase of p-Akt was produced by cocaine and 7-OH-DPAT in cultures from wild-type mice, while no effect was observed in mesencephalic neurons from D3KO mice, as supported by two-way anova significant genotype effect [F(1,18) = 35.81, p < 0.0001], treatment effect [F(2,18) = 8.33, p < 0.005], and ‘treatment × genotype’ interaction [F(1,14) = 9.23; p < 0.0001] (Fig. 7b).
Inhibition of PI3K in primary mesencephalic cultures blocks cocaine-dependent ERK and Akt pathway activation and prevents structural plasticity
The involvement of PI3K-ERK and -Akt pathways in dendrite outgrowth and soma size changes produced by cocaine was further studied using the PI3K inhibitor LY294002. Twenty minutes pre-treatment with LY294002 (10 μM) antagonized the significant increase of p-ERK levels [F(3,16) = 9.71; p < 0.01] (Fig. 8a) and p-Akt levels [F(3,16) = 49.76; p < 0.001] (Fig. 8b) measured by western blot in mesencephalic cultures 2 min after the challenge with cocaine (10 μM). Similar results were obtained in parallel experiments using double-staining immunofluorescence aimed to localize the expression of either p-ERK-IR or p-Akt-IR in dopaminergic neurons. Pre-treatment with LY294002 completely antagonized cocaine-induced p-ERK-IR (Fig. 8c) and p-Akt-IR (Fig. 8d) in dopaminergic neurons at 2 min after challenge. No effects on ERK or Akt phosphorylation were observed with LY294002 incubation alone (Fig. 8c and d).
The effects on cocaine-induced structural plasticity of the same dose of LY294002 used to block ERK and Akt phosphorylation were investigated at 72h after exposure to cocaine. LY294002 significantly attenuated cocaine effects on dendrite outgrowth and soma area (Fig. 8e–g). Accordingly, when maximal dendrite length was measured, two-way anova indicated significant cocaine effect [F(1,116) = 4,00; p < 0.05] and interaction [F(1,112) = 6,71; p < 0.02], the post hoc test showing LY294002 attenuating the effect of cocaine (Bonferroni’s test, p < 0.05) (Fig. 8e). When data from number of primary dendrites were investigated, two-way anova indicated significant cocaine effect [F(1,116) = 4,47; p < 0.05] and interaction [F(1,112) = 3,89; p < 0.05], the post hoc test showing again the LY294002 blocking effect on the response to cocaine (Bonferroni’s test, p < 0.01) (Fig. 8f). Similar result were observed for the soma areas, two-way anova showing highly significant cocaine effect [F(1,116) = 13.06; p < 0.0001] and interaction [F(1,112) = 30.55; p < 0.001]. The post hoc test showed significant attenuation of cocaine effects by LY294002 (Bonferroni’s test, p < 0.001) (Fig. 8g). For all measurements, no effect was observed when LY294002 was tested alone (Fig. 8e–g).
The present study demonstrates that long-term exposure to cocaine increased soma area and dendrite arborisation of mesencephalic dopaminergic neurons via a mechanism dependent on dopamine D3R activation. These morphological changes were detected both in vitro, using primary neuronal culture preparations from E12.5 mouse embryo mesencephalon, and in vivo, in the VTA and SN of P1 offspring prenatally exposed to cocaine during the same embryogenic period. The in vitro studies were conducted using a validated protocol of enriched primary neuronal cultures containing 5–7% TH-IR or DAT-IR neurons and less than 0.5% astrocytes (Collo et al. 2008). The maximal dendrite length and the number of primary dendrite were significantly increased 24 hours after cocaine exposure, while no significant effect was produced on the soma area, the proper increase occurring only at 72 h. The delayed effect in soma was probably related to differences in the geometry of soma surface-volume versus dendrites, and to the different molecular substrates associated with their growth: apposition of new membrane and development of Golgi outposts for the remodelling of primary dendrites (Horton et al. 2005), and changes of neurofilament proteins and of global axonal trafficking for the growth of the soma size (Nestler et al. 1993; Russo et al. 2007).
The morphological changes produced by 1–10 μM cocaine were paralleled by a significant increase of the [3H]-DA uptake, indicating an augmented capacity of synaptic, dendritic and somatic DA transport. Interestingly, up-regulation of DAT was reported in the VTA and SN of rodents and monkeys exposed to cocaine in utero (Fang and Ronnekleiv 1999; Glatt et al. 2000). Increases of dendrite arborization, soma area and DA uptake in mesencephalic neurons were previously observed following exposure to 1–10 μM amphetamine (Mueller et al. 2006; Collo et al. 2008), another psychostimulant that increases extracellular DA levels (Rouge’-Pont et al. 1999).
In the present work, we investigated the relevance of specific subtypes of DA receptors that could be engaged by the increased extracellular DA levels produced by cocaine at first using pharmacologic antagonists and then knockout mice. An involvement of D1R was excluded by showing that the selective antagonist SCH23390 was inactive, consistent with the known absence of expression of D1R in mesencephalic dopaminergic neurons. Evidence for an involvement of D2R and/or D3R was obtained with sulpiride, which possesses similar affinity for both receptors: sulpiride suppressed the increases of dendrite arborization and soma area produced by cocaine. These effects were reproduced with the preferential D3R antagonist SB-277011-A at concentration close to its affinity for D3R (50 nM; Reavill et al. 2000) and confirmed using the more potent and selective D3R antagonist S-33084 (Millan et al. 2000) at concentration as low as 2 nM. These results were in line with the blocking effect of low dose SB-277011-A on the morphological changes generated in mesencephalic dopaminergic neurons by amphetamine (Collo et al. 2008). Interestingly, in the present work we showed that pre-treatments with SB-277011-A did not block the increase of dendrite outgrowth and soma size produced by the neurothrophin BDNF that engages a different receptor but similar intracellular signalling pathways (Goggi et al. 2003).
The critical role of D3R for structural plasticity induced by cocaine was confirmed by testing mesencephalic cultures prepared from D3KO mice embryos. Dopaminergic neurons from D3KO mice did not respond to cocaine and 7-OH-DPAT, a selective D3R antagonist, while neurons from wild-type mice did. Interestingly, BDNF was effective increasing dendrite arborisation and soma size in both D3KO and wild-type dopaminergic neurons, suggesting that the neurotrophin-operated signalling was indeed functional in D3KO mice.
Independent support for a role of D3R was obtained by the exploration of intracellular pathways activated by D3R via PI3K in cell lines, which results in phosphorylation of ERK and Akt (Cussac et al. 1999; Beom et al. 2004; Mannoury la Cour et al. 2011) and other downstream genes known to participate to cytoskeletal regulations (Qian et al. 2004). More recently, using primary cultures of mesencephalic dopaminergic neurons, we showed that LY294002 blocked D3R agonist-dependent ERK phosphorylation within minutes, and D3R agonist-dependent dendrite outgrowth after 72 h incubation (Collo et al. 2008). In the present work, we showed that exposure to cocaine produced a rapid phosphorylation of both ERK and Akt peaking after 2 min, increasing dendrite outgrowth and soma area after 72 h incubation. Cellular localization within dopaminergic neurons was supported by double-staining immunofluorescence obtained with antibodies recognizing dopaminergic markers (TH or DAT) and the phosphorylated kinases (p-ERK, p-Akt), respectively. These data are in line with evidence of psychostimulants increasing ERK phosphorylation in vivo, in particular in the VTA (Dietz et al. 2009), and producing different effects on PI3K phosphorylation depending to the brain regions (Zhang et al. 2006). To our knowledge, the present data are the first showing a critical role for PI3K/Akt activation in the structural plasticity effects of cocaine in dopaminergic neurons of the mesencephalon. The direct involvement of PI3K/Akt and ERK in dendrite outgrowth was previously shown in hippocampal neurons (Jaworski et al. 2005; Kumar et al. 2005), while its role of regulator of soma area was described in a variety of neurons (Scheidenhelm et al. 2005). In particular, genetic or pharmacologic manipulations aimed to reduce or increase activity of PI3K/Akt in dopaminergic neurons were associated with soma area reduction (Russo et al. 2007) and enhancement (Domanskyi et al. 2011; Kim et al. 2011), respectively. In the present work, we describe soma area increase in mesencephalic dopaminergic neurons of wild-type newborn mice after in utero exposure to cocaine. We also showed that this effect was absent in D3KO newborn mice, corroborating the observations made in primary cultures and supporting the need of a viable D3R to mediate morphological effects of cocaine. Structural plasticity depending on prenatal cocaine exposure was consistently described in prefrontal cortical neurons (Morrow et al. 2007; Lu et al. 2009); our data open to the possibility that changes in mesencephalic neurons could participate to the long-lasting behavioral liabilities to psychostimulants in adult life (Crozatier et al. 2003). However, more data are required to support this view. In this study, performed on newborns at P1, the prenatal effects of cocaine were observed in the dopaminergic neurons of both VTA and SN. This result was at variance with observations made in published studies on young adults (P-30-45 or older), when cocaine changes are restricted to VTA (Mueller et al. 2006; Russo et al. 2007; Sarti et al. 2007; Spiga et al. 2010), suggesting dependency on developmental stage.
Overall, the present investigation provided the first direct evidence that cocaine increases dendritic arborization and soma size in mesencephalic dopaminergic neurons, a phenomenon that could contribute to the maladaptive structural plasticity that is believe to constitute a pathogenetic drive for addiction (Nestler et al. 1993; Robinson and Kolb 2004; Sarti et al. 2007; Russo et al. 2010). Using in vitro and in vivo protocols, pharmacologic tools, knockout mice and intracellular signaling we provided evidence supporting a critical role of the DA D3R in mediating cocaine-dependent structural plasticity in mesencephalic dopaminergic neurons. A robust body of pre-clinical data supports a role for D3R in psychostimulant addiction, with D3R antagonists preventing relapse to drug-taking in abstinent individuals (Caine and Koob 1993; Pilla et al. 1999; Sokoloff et al. 2001; Heidbreder et al. 2005). Therefore, it is possible that D3R antagonists operate, among other mechanisms, via modifications of cocaine-induced structural plasticity in the mesencephalic dopaminergic neurons. Recent evidence obtained with Positron Emission Tomography indicates that the most abundant localization of D3R in humans brain is in the ventral mesencephalon, where VTA and SN are located (Searle et al. 2010). These data further support the idea that targeting dopaminergic neurons with D3R medicaments could provide novel therapeutics approach to psychostimulant addiction.
The authors thank Marzia Melotti for technical support. This study was supported by grant from ex 60%, University of Brescia to GC, and by grant from MIUR (PRIN 2008) to CM. EMP is full-time employee of GlaxoSmithKline. MJM. is full time employee of Institut de Recherches Servier. The other authors declare no conflict of interest in this work.